Simming HiLo vs KO

[bj21abc]

Hi,

Was wondering if anyone has any experience simming HiLo Vs KO ?

I have been simming a (6D S17 75%pen ES DAS 1-20 spread) game with both - using 1/2 kelly betting ramps which I assume are more accurate for KO (more granular).
For HiLo TC conversion I use round to full decks and floor. Am using CV Data, with hilo and KO indices which have been derived (are quite close to standard).

Since KO is supposed to slightly edge out HiLo in both BE and PE, I was expecting some difference - but the results (going up to 2,000 mn rounds) show almost identical SCORE. Would be glad to hear if anyone has carried out similar sims with different results...

Thanks

D.

 

[QFIT]

Nearly identical is what you should expect.

 

[FLASH1296]

Nearly identical is what I would have expected.

I would have been quite concerned if it wasn't.

It is not like you are looking at simulating a level two count.

Do yourself a really big favor.

Read QFIT's "Modern Blackjack"

 

[bj21abc]

I would expect close, but not identical.
With the BE (.98 vs .97) and PE differences (.55 vs .51) I was expecting a 1% or so increase in SCORE, and 20,000 mn rounds should be enough to see that difference.

I could be doing something wrong, but would be interested to see if this has been done before (must have been...) with different results.

D.

 

[FLASH1296]

P. E. is NOT crucial in shoe games, as they are in "pitch" games.

 

[bj21abc]

Not "as crucial" but still quite important.

Don't have the figures to hand but as I recall for a 6D game (play all) strategy variation makes up ~ 20-25% of the gain, with betting variation (and a little bit of insurance) making up the rest.

This is why with the given differences in BC and PE above, I am assuming there should be a corresponding additional ~ 1% advantage.

 

[zengrifter]

Since KO is supposed to slightly edge out HiLo in both BE and PE, I was expecting some difference...

Who said that KO should edge out HiLo? zg

 

[QFIT]

Not "as crucial" but still quite important.

Don't have the figures to hand but as I recall for a 6D game (play all) strategy variation makes up ~ 20-25% of the gain, with betting variation (and a little bit of insurance) making up the rest.

This is why with the given differences in BC and PE above, I am assuming there should be a corresponding additional ~ 1% advantage.

Well, two points. First, PE is only an estimate. Secondly, it doesn't actually apply to running count strategies.

 

[bj21abc]

Why would PE not apply to RC strategies ? If all point values do not add up to zero then does the set formula not apply ?

And if so, why does KO always have a PE quoted ?

 

[StandardDeviant]

It's funny. We sweat the differences of one system versus another. Yet these differences only have meaning in the long run, which none of us will live to see. The lesson is pretty obvious...go with the system you are most comfortable using and hope for the best. If you believe in your system you will continue through the downturns. If you don't believe, you are, in a word, finished. (If you want, choose another f-word.)

 

[QFIT]

Read Theory of Blackjack. PE was designed for balanced strategies. And the commonly used EORs are specifically aimed at single-deck with a huge number of indexes.

That said, the numbers are surprisingly useful for other circumstances.

 

[QFIT]

It's funny. We sweat the differences of one system versus another. Yet these differences only have meaning in the long run, which none of us will live to see. The lesson is pretty obvious...go with the system you are most comfortable using and hope for the best. If you believe in your system you will continue through the downturns. If you don't believe, you are, in a word, finished. (If you want, choose another f-word.)

Agree and not. On average, we see the long term results. Otherwise, we would all smoke and eat ice cream sundaes all day.

 

[Brock Windsor]

Run the sim with a smaller spread (8:1) and compare that result to the 20:1 spread result.
BW

Not "as crucial" but still quite important.

Don't have the figures to hand but as I recall for a 6D game (play all) strategy variation makes up ~ 20-25% of the gain, with betting variation (and a little bit of insurance) making up the rest.

This is why with the given differences in BC and PE above, I am assuming there should be a corresponding additional ~ 1% advantage.

 

[k_c]

Count Comparison Project

I am working on an approach to comparing counting systems. It uses basic strategy (no indices) so the only difference in the results for each count comes from their betting spreads.

Basically I sim a number of shoes to an input penetration and record some basic data:
1. shoe's composition prior to each round
2. expected value using basic strategy prior to each round using combinatorial analysis
3. whether player, dealer or both got blackjack on the round
4. result of round using basic strategy

From the above data most any count can be charted for whatever spread is chosen.

Right now the vehicle I'm using to display results is an Excel spreadsheet. I've made the spreadsheet open source so anyone can add a new feature. (Dead link: http://www.bjstrat.net/simDemo.html) _Project demo_ can be dowmloaded here at the bottom of page.

Basically the process is to import data into the spreadsheet using a macro built into the sheet named ImportData. Recently I got I new computer which generates data faster so I was able to include a sample data file with more data. Once data is input spreads can be changed as desired. The new spreads will be reflected in the spreadsheet. Differing counts can be displayed and the sheet updated as well.

Also I revised my shuffle algorithm for the program that generates the data. If anyone already has the spreadsheet, the sheet can be restored to an empty state by running the DeleteAllData macro. Data using the revised shuffle can then be imported from the data file zip attached to this post.

I would be interested in feedback on the project.